IBM Mainframe Forum

Hi All,

I would like to know how to calculate the Pic size of a COMP-3 and COMP variables.
My requirement is in my COBOL program one variable is declared as S9(9)V99 COMP-3.Now i would like to declare an equivalent alpha numeric for the COMP-3 variable. wat is the Pic size of alpha numeric ( X(??)) variable.
How to calculate the size of equivalent alpha numeric for COMP-3 and also COMP ??

Thanks in advance.

regards,
Prashanth

For COMP-3, it is simple..... Add up the number of 9s, add one for the sign and then divide by two and round up.....
S9(9)V99 is X'99999999999S' or six bytes
S9(8)V99 is X'09999999999S' or six bytes
S9(7)V99 is X'999999999S' or five bytes

For COMP, I suggest that you look it up yourself.
Pay attention to both the COMP (binary) and the COMP-5 (native binary)....

Hello,

Now i would like to declare an equivalent alpha numeric for the COMP-3 variable.

Depending on exactly what you want to do, knowing the physical length of the comp/comp-3 variables may not be all that you need to know.

If you explain how you want to use the "equivalent alpha numeric", we may be able to offer more usable suggestions.

hi

the comp-3 values can be calculated by using the formula

(n/2+1)(when the number is even) for instance pic s9(4) comp-3 ,so here n is 4 by using the formula we get 3 bytes.
(n/2 +1/2) (when the number is odd) for instance pic s9(5) comp-3 sh here n is 5 and answer is also 3 bytes.

and for comp s9(1) to s9(4) - 2 bytes
s9(5)to s9(9)- 4bytes
s9(10)to s9(18) 8bytes

These users thanked the author suryaprakesh for the post:

rajeshdatrika (Wed Feb 15, 2017 3:44 pm)

prashanthreddy wrote:Hi All,

I would like to know how to calculate the Pic size of a COMP-3 and COMP variables.
My requirement is in my COBOL program one variable is declared as S9(9)V99 COMP-3.Now i would like to declare an equivalent alpha numeric for the COMP-3 variable. wat is the Pic size of alpha numeric ( X(??)) variable.
How to calculate the size of equivalent alpha numeric for COMP-3 and also COMP ??

Thanks in advance.

regards,
Prashanth

None of the mentioned answers are correct.
The question is declare an equivalent PIC X variable that should be 9 + 2 = PIC X (11).

The calculation which is being done in the other replies is about the bytes compiler takes to store the comp-3 variable internally to save memory.

Seniors correct me if I am wrong

sinmani , since the original question was posted more than 4 YEARS AGO, do you really think your answer will make a difference to that person?

Actually 9+2+1 for the decimal point (if you want to display that as well)

Robert Sample wrote:sinmani , since the original question was posted more than 4 YEARS AGO, do you really think your answer will make a difference to that person?

Hi Robert ,

It might not help the person who asked that question but it can definitely help thousands of readers who flock to this forum for gaining knowledge.

And since the original question was so poorly worded that we don't know what was wanted... I can't exactly say it wasn't a request to know the bytes rather than the digits...

i will make you understand with examples-
formula is - [(N/2)+1] = and Truncate it to Last pure number
Eg- S9(11) = [(11/2)+1] = 6.5 --> round up to last pure no which is 6
OR
S -takes 1/2 byte and N/2
S9(7) = 1/2+7/2 = 4
S9(8) = 1/2+8/2=4.5 = 5 (NEXT PURE NUMBER)

S9(7)V99 COMP 3 = 7+2=9 S9(9) = 1/2+9/2=5 BYTE